In the optical communications space, various techniques are used to synthesize an optical communications signal for transmission. A popular technique utilizes a laser 2 coupled to an external optical modulator 4, as shown in FIG. 1. The laser 2 generates a narrow-band continuous wave (CW) optical carrier signal 6 having a desired wavelength. The optical modulator 4 operates to modulate the amplitude and/or phase the carrier signal 6 to generate the optical communications signal 8 based on a drive signal 10 that encodes data to be transmitted. Typically, the drive signal 10 is generated by a driver circuit 12, which normally provides a power amplifier for amplifying the power of an input digital data signal x(m) to satisfy the input power requirements of the modulator 4.
In the arrangement illustrated in the FIG. 1, the optical modulator 4 is provided by a well known Mach-Zehnder (MZ) interferometer. Other types of modulators may be used, depending on the desired type of modulation. For example, an electro-absorptive modulator (EAM) may be used for amplitude modulation; whereas phase modulators are well known for implementing phase modulation. In each case, the driver circuit 12 generates the drive signal 10 by scaling the input data signal x(t) to satisfy the voltage and current requirements of the modulator 4. The driver circuit 12 may also generate one or more bias signals (not shown) for controlling a bias point of the modulator 4 in a manner well known in the art. The amount of modulation achieved is proportional to the product of effective length of the modulator and the drive voltage applied. The length of the modulation element is fixed by the construction of the modulator, and the drive voltage is adapted to achieve the desired amount of modulation. Control circuits are generally used to achieve this adaptation.
FIG. 2 illustrates an alternative arrangement known from Applicant's co-pending U.S. patent application Ser. No. 10/677,223 filed Oct. 3, 2003. In that system, a complex driver circuit 14 comprises a digital filter 16 which uses the input data signal x(m) and a compensation function c(t) to calculate multi-bit In-Phase and Quadrature component values I(n) and Q(n) of a target optical E-field modulation. A non-linear compensator 18 uses the I(n) and Q(n) components to compute multi-bit sample streams VR(n) and VL(n). These digital sample streams are then converted into corresponding analog voltage levels by respective multi-bit digital-to-analog converters (DACS) 20, filtered (at 22) to reduce out-of-band noise, and scaled by low noise amplifiers 24 to yield a pair of drive signals VR(t) and VL(t) which are supplied to respective branches of the MZ modulator 4. If desired, respective digital filters (not shown) may be positioned between the non-linear compensator 18 and the DACs 20 in order to compensate any propagation delay differences between the DACs 20 and the MZ modulator 4.
As may be seen in FIG. 3a, an electro-optic component typically includes a control region 26 defined by a pair of electrodes 28 placed on opposite sides of an optical waveguide 30. With this arrangement, an electric field through the waveguide material can be set up by applying a voltage across the two electrodes 28. For High-speed applications (i.e. for micro-wave frequency drive signals), the electrodes 28 are typically implemented using strip-line techniques, with the drive signal v(t) supplied to the up-stream end of the electrode 28, relative to the direction of propagation of light through the waveguide 30. The opposite (downstream) end of the electrode 28 is typically terminated by a matched resistive load to ground (not shown) to prevent unwanted signal reflections. Ideally, the characteristic impedance of the electrode 28 is selected such that the drive signal propagates through the electrode 28 at the same speed as the light propagating in the optical waveguide 30.
The arrangement shown in FIG. 3a is typical of a voltage-controlled phase modulator. In such a phase modulator, the refractive index of the waveguide is a function of the applied voltage, so that light propagating though the control region 26 will experience a phase delay that is proportional to the applied voltage and the length L of the electrodes 28. In order to maximize the phase delay, the drive signal V(t) is commonly supplied as a differential signal pair ±V(t), which doubles the magnitude of the voltage across the electrodes 28.
FIG. 3b shows a dual-branch MZ modulator 4 constructed using a pair of phase modulators of the type illustrated in FIG. 3a. For simplicity, bias control circuits which are normally provided as part of the MZ modulator 4 are not shown. Voltage inverters 32 in each drive signal path convert the applied drive signals Vx(t) into corresponding differential voltage pairs ±Vx(t). If desired, the drive signals Vx(t) can be generated by the complex driver 14 described above with reference to FIG. 2.
The arrangement of FIGS. 2-3b is particularly advantageous in that the multi-bit sample values VR(n) and VL(n) can be computed taking into account non-linearities of the analog signal path (e.g. the DACs 20, filters 22 and LNAs 34) and the MZ modulator 4, such that the optical E-field of the composite signal 8 appearing at the output of the MZ modulator 4 closely matches the target E-field modulation computed by the digital filter 16. In some embodiments, the compensation function c(t) is selected to compensate impairments of an optical link not shown, in which case the target E-field modulation represents a pre-distorted signal which will be transformed by the link impairments into a substantially undistorted optical signal at a receiver end of the link.
As will be appreciated, at least the digital filter 16, non-linear compensator 18 and DACs 20 of the complex driver 14 can be implemented on a single Application Specific Integrated Circuit (ASIC). This arrangement provides advantages in terms of performance, power consumption and cost. For example, in some embodiments, an ASIC implemented using Complementary Metal Oxide Semiconductor (CMOS) technology can cost-effectively generate the analog DAC output signals to 6-bits precision at a sample rate of 20 GHz. This performance is sufficient to compensate even severe impairments of the optical link at data rates of 10 Gb/s.
However, electro-optical components such as MZ modulators are typically fabricated using techniques that are not readily compatible with those of integrated circuits (ICs). In many cases, the materials used for electro-optical components differ from those used in IC fabrication, or from the preferred technology (e.g. CMOS, GaAs etc.) of the driver IC, which would require redesign of the IC production line. Even where this is done, the drive signals V(t) required to obtain satisfactory dynamic range of the electro-optical component may significantly exceed the voltage and/or thermal limits of the IC. For example, in the system of FIG. 2, each of the drive signals VR(t) and VL(t) may require a power level on the order of 10 Watts or more.
In view of the above difficulties, the driver 12 or 14 and the optical modulator 4 are typically provided as separate packages mounted on a printed circuit board, separated by the signal conditioning components (e.g. filters and amplifiers) required to modify the drive signals output from the driver ASIC to satisfy the input power requirements of the optical modulator 4. In many cases, this is a satisfactory arrangement. However, in some applications it is desired to provide a smaller, lower power assembly.
U.S. Pat. No. 4,288,785 (Papuchon et al) teaches a digitally controlled light intensity (amplitude) modulator which is compatible with Transistor-Transistor Logic (TTL). According to Papuchon et al, each bit of an N-bit control word is used to control the voltage supplied to a respective electrode of an MZ modulator. In one embodiment, the electrodes are arranged symmetrically on both branches of the MZ modulator, and receive the same control signal. In other cases, the electrodes are arranged on one branch of the MZ modulator, while the other branch receives a feed-back signal designed to linearize the MZ modulator response. In some embodiments, the length of each electrode varies in a geometric scale according to the binary weight of its control bit. In other embodiments, each electrode has the same length, and voltage dividers are used to scale the voltage supplied to each electrode, again according to the binary weight of its control bit.
It will be appreciated that the modulator of Papuchon et al is a high voltage device designed to operate at relatively low speeds. In particular, the modulator is driven by TTL logic, which produces a 5 volt swing between binary ‘0’ and ‘1’ logic states. It is also well known that TTL logic is a relatively high-current and low yield technology, which precludes its use in large integrated circuits with clock speeds higher than about 106 μHz. However, high speed integrated circuits, for example high speed CMOS IC's operating at clock speeds exceeding 109 Hz, typically have a voltage swing on the order to 1 volt, and are capable of sourcing only very limited currents. This leads back to an assembly in which the driver IC and the optical modulator 4 must be provided as separate packages mounted on a printed circuit board, separated by the signal conditioning components (in this case high voltage drivers). In addition, the feedback loop used to linearize the sinusoidal response of the MZ modulator includes an inherent loop delay. This loop delay imposes an upper limit on the speed of the control word, beyond which the MZ modulator response will become increasingly non-linear.
Because Papuchon is considering low speed modulations with an idealized MZ modulator, he does not indicate methods to mitigate or avoid the nonidealities of an actual modulator operating at high speeds, with physical limitations and imperfect manufacturing.
Digitally driven optical modulators capable of high-speed operation are known, for example, from A 12.5 GSample/s Optical Digital-to-Analog Converter with 3.8 Effective Bits, A. Leven et al., Lasers and Electro-Optics Society, 2004, and references cited therein. In each case, multiple parallel phase modulators are provided, each of which is controlled by a respective drive voltage vi, which appears to be scaled to follow the binary weight of a respective bit of a digital control signal so as to impose a phase shift proportional to the binary weight of that bit. The signals emerging from the parallel phase modulators are coherently combined to yield a composite amplitude modulated optical signal, the intensity of which is proportional to the value of the digital control signal.
An alternative arrangement is described in Digital-to-analog Conversion Using Electrooptic Modulators, Yacoubian et al, IEEE Photonics Technology Letters, Vol. 15, No. 1, January 2003. In this case, multiple parallel amplitude modulators are provided. Each modulator is driven by a respective bit of a digital control signal, and attenuates light in response to the logic state of that bit. Binary-scale weighting of each amplitude modulator is obtained by using a weighted 1-to-N coupler which divides a continuous wave (CW) carrier to produce the branch signals supplied to each controlled amplitude modulator. The signals emerging from the parallel amplitude modulators are in-coherently combined on a photodetector to yield an electrical signal having an amplitude that is proportional to the value of the digital control signal.
Unlike the system of Papuchon et al, the systems of Leven et al. and Yacoubian et al are capable of operating at high speeds. The system of Yacoubian et al is also capable of operating with sub-1 volt control signals, and thus cold be driven by a high speed CMOS IC, for example. However, neither of these systems is suitable for an optical synthesizer in which the driver IC and optical modulator are integrated within a common package. In particular, the system of Leven requires a drive circuit capable of delivering a scaled voltage to each phase modulator, which precludes direct connections between the phase modulators and a driver IC. In the case of Yacoubian, the signals emerging from the optical modulators must be recombined incoherently in order to avoid errors. This precludes operation as an optical synthesizer, because it cannot reliably generate a low-error optical signal.
Accordingly, methods and apparatus enabling a high speed driver IC and an optical modulator to be integrated within a common package remain highly desirable.